libsForQt5.qgpgme fails tests (breaking kde/plasma)

[hedning]

Issue description

qgpgme failure on hydra (this is a qt specific test):

FAIL! : TofuInfoTest::testTofuConflict() 'sig.validity() == Signature::Marginal' returned FALSE. ()
Loc: [t-tofuinfo.cpp(458)]
PASS : TofuInfoTest::cleanupTestCase()
Totals: 7 passed, 1 failed, 0 skipped, 0 blacklisted, 1047ms

kde breakage

This seems to affect master too, but for some reason it's been built successfully once.

The aarch64 build started failing when updating qt5 from 5.11 to 5.12 making it the likely root cause: https://hydra.nixos.org/build/86769014

Edit: the cause was expired testing keys: https://dev.gnupg.org/T3815#121469

cc gpgpme maintainer @primeos
cc @bkchr as you've been active in keeping plasma/kde up to date lately.

ping staging next: #53483

[bkchr]

Did you found this? Sounds very similar and should be fixed...

[hedning]

No didn't find that, but yeah, that sure looks exactly like the same issue. I'm guessing the qt issue were never actually fixed:

That leaves the qt issue to be done.

Blame on the test gives no change at least:
https://github.com/gpg/gpgme/blame/b182838f71d8349d6cd7be9ecfb859b893d09127/lang/qt/tests/t-tofuinfo.cpp#L458

[hedning]

And applying the suggested patch does fix the test:

+++ gpgme-1.10.0/lang/qt/tests/t-tofuinfo.cpp
@@ -455,7 +455,7 @@ private Q_SLOTS:
 
         QVERIFY(result.numSignatures() == 1);
         auto sig = result.signatures()[0];
-        QVERIFY(sig.validity() == Signature::Marginal); 
+        QVERIFY(sig.validity() == Signature::Marginal || sig.validity() == Signature::Unknown);
 
         auto stats = sig.key().userID(0).tofuInfo();
         QVERIFY(!stats.isNull());

[bkchr]

So, what are we want to do? Do you maybe want to comment on the upstream issue?

[hedning]

Hmm, strange, I'm getting the same failure on master (using nix-build --check), forcing single core does not fix the problem though, so probably not a race condition.

I've reported the failure as you suggested: https://dev.gnupg.org/T3815#121448

[hedning]

The cause is most likely qt 5.11 to 5.12 upgrade, as the aarch64 build started failing the test then: https://hydra.nixos.org/build/86769014

[hedning]

So the keys used in this test expired on 2019-01-06, which explains why we suddenly saw this without any apparent reason. https://dev.gnupg.org/T3815#121469 :

Oh,.. it is even worse. The conflict keys expired 2019-01-06 so they are actually expired right now.

I'll change the test to use keys without expiry.

[primeos]

@hedning Thanks for fixing this and sorry for my late reply...

(Edit: And of course also thanks @bkchr)